Hi-res Saturn rings - inspired by Ciclops picts

[Boux]

Hi there!
I am making my own Saturn rings high-res 4k textures.
The basic template comes from the Ciclops website.
I have aligned the texture with the sheperd moons orbits as precisely as possible.
The transparency profile has been benchmarked against existing ring textures.
It looks best after commenting out the "Color" statement in the Saturn rings block of solarsys.ssc.

If you want to test the texture, just go grab it there:
http://jmmi.club.fr/celestia/ciclops/saturn_rings_ciclops-4k.png
Comments are welcome

[t00fri]

What did you chose for the transparency profile across the rings? It is known to vary strongly from earthbound observations of the transmitted light of a star passing across the ring!

Bye Fridger

[Boux]

As said, I checked against existing Celestia ring textures to make sure this one is close to current standards. There is a lot of work in it.
There is an interesting discussion there:
http://www.planetary.org/news/2005/0913_Changing_Views_of_Saturns_Rings.html
I specially like the inter-title "It All Depends on Your Point of View".
Unfortunately, the Celestia engine is unable to render properly the rings because they have no thickness and they are not rendered by a particle system or discrete objects. The texture is flat and static.
Celestia can only render an approximation. You know that of course

[t00fri]

As said, I checked against existing Celestia ring textures to make sure this one is close to current standards. There is a lot of work in it.
There is an interesting discussion there:
http://www.planetary.org/news/2005/0913_Changing_Views_of_Saturns_Rings.html
I specially like the inter-title "It All Depends on Your Point of View".
Unfortunately, the Celestia engine is unable to render properly the rings because they have no thickness and they are not rendered by a particle system or discrete objects. The texture is flat and static.
Celestia can only render an approximation. You know that of course

Boux,

but the transparency profile that we used for the Saturn ring from the best available publication is a /complicated, strongly varying function/ of the radial distance. This you can impossibly guess. If you did not copy it into the alpha channel of your texture it's not going to be right.

How does the alpha channel of your ring looks like???

Bye Fridger

[Boux]

I am still working on it.
Looking at these profiles:
http://www.mmedia.is/~bjj/data/s_rings/transparency.png
http://www.lowell.edu/users/spencer/satringprof2.bmp
Well, for one zillion reasons, the exercise is purely academic.

[Boux]

OK, here is the final version of my hires Saturn rings.
It should be fine now as the transparency profile is in line with published data.
It is a crazy job to do that. I probably never do it again.
t00fri should be happy now
First, a new screenshot to give an idea:


Then, the cel url to go right there in Celestia:
cel://follow/Saturn/2005-12-28T09:16:45 ... 07159&lm=0

And of course the link to the texture for you guys to give a try:
http://jmmi.club.fr/celestia/ciclops/sa ... nal-4k.png
Enjoy. Feedback welcome.

[selden]

I just thought I'd remind you that 4K Ring textures only work with Nvidia 4nnn series cards or better. ATI cards are limited to 2K Ring textures and many older cards are limited to 1K Ring textures.

And, for what it's worth, Grant Hutchison's 4K Saturn Ring texture is available at http://www.lepp.cornell.edu/~seb/celest ... rings.html

[Boux]

Hi Selden,
It is easy to downsize the texture to lower resolutions. If somebody needs a lower-res version, I will do it.
I think there is no point in doing this though as the one bundled with Celestia is OK for the average PC.
Grant's texture is the same as the bundled one and is pretty old (7Jul 04).

[Boux]

Hi all,
This morning, I noticed that I had uploaded the wrong picture and texture
I messed up the couple dozens working layers and uploaded a bad one with empty placeholders intended for detail work.
Anyway here are the very final ones in 4, 2 and 1k sizes so that any hardware will be able to render.
http://jmmi.club.fr/celestia/rings-30-12/saturn_rings_ciclops_final-4k-12-3005.png
http://jmmi.club.fr/celestia/rings-30-12/saturn_rings_ciclops_final-2k-12-3005.png
http://jmmi.club.fr/celestia/rings-30-12/saturn_rings_ciclops_final-1k-12-3005.png
And a new screenshot:

[t00fri]

Boux,

I have extracted the alpha channel from your latest rings
(4k) and compared it with the one that was derived directly
from the earth-bound hires star passage data in Grant's 4k
version:

OK, I've downloaded and converted the ring occultation
data at http://ringmaster.arc.nasa.gov/cds/EBROCC_0001/.
So I've now got eight separate alpha channels of varying
quality, with some data dropouts, each consisting of a
6900-pixel strip spanning radial distances of 73004.5km to
141994.5km from Saturn's centre, at 10km resolution.

I find that your alpha profile is in disagreement with these
direct transparency measurements
http://pds-rings.arc.nasa.gov/ringocc/ringocc.html
in many places...To syncronize and merge these transparency data with the albedo imaging from the Cassini
mission was a tedious piece of work. It's present 1k version is part of the official Celestia distribution.

Bye Fridger

[Boux]

Yes, it is tedious to say the less
Stronger words are coming to my mind.
Total accuracy would need a pixel row by pixel row work on the transparency.
Even after corrective transformations of the Ciclops picture, the transparency layer has to be adjusted row by row. There are remaining perspective deformations.
The ideal shot would have to be taken from a distance, perpendicularly to the middle of the rings plane.

[t00fri]

Yes, it is tedious to say the less
Stronger words are coming to my mind.
Total accuracy would need a pixel row by pixel row work on
the transparency.
Even after corrective transformations of the Ciclops picture,
the transparency layer has to be adjusted row by row.
There are remaining perspective deformations.
The ideal shot would have to be taken from a distance,
perpendicularly to the middle of the rings plane.

Boux,

unfortunatleyl I don't understand at all what you are
describing.

Let me summarize a few issues:

In all cases the approximation is to assume that both the
ring albedo and the ring transparency are
only functions of the /radial/ distance/, hence we have
effectively only 2 line profiles to map onto each
other! In practice, we took a 2 pixel height of each line. But
for working and inspecting one my temporarily scale it up
to 100 or 200 pixels height (cf the image below). The
transparency data from the given references may be
directly read into such a profile and displayed. I suppose
you did this, otherwise there would be no chance for you to
get the transparency right. I did not understand though why
your alpha channel disagrees with those data?

Here is the 1k version (scaled to 800 pix) but we do have
higher resolutions as well, of course.


As usual for alpha channel data, white is completely untransparent, while black corresponds to complete transparency.

So the main work consists in /distance/ normalizing the albedo profiles correctly to the transparency profile the
values of which are directly tabulated in [km]. Perhaps I
can understand better if you explain what your REALLY did...

Happy New Year!
Bye Fridger

[fsgregs]

Hi guys:

I tried uploading Boux's ring textures today, but the links no longer work. Anyway, that leads to a new question. Recent published photos reveal that both Saturn and Uranus have a decidedly bluish outer ring. Is there any plans to redo the rings for Saturn and Uranus to include it?


Frank

[buggs_moran]

I just got one from 4 posts up just fine. Maybe the server was down.

[PeterC]

On my screen of Celestia, the Saturn's rings are not good qualities (on the left of the big image). I tried to add the new rings with circle perfect (on the right of the big image) but Saturn and rings (little image below) are very bad on Celestia for that.



Low-resolutons = 2 x 1024 pixels
I cut an image of Boux to show you. High-resolutions = 2 x 4096 pixels

How to add the rings high resolution? Please light me precisely.

[Boux]

On my screen of Celestia, the Saturn's rings are not good qualities (on the left of the big image). I tried to add the new rings with circle perfect (on the right of the big image) but Saturn and rings (little image below) are very bad on Celestia for that.



Low-resolutons = 2 x 1024 pixels
I cut an image of Boux to show you. High-resolutions = 2 x 4096 pixels

How to add the rings high resolution? Please light me precisely.

Hi,
Just edit your celestia.cfg file and change the sample point and sections values as follows:

Code: Select all

#------------------------------------------------------------------------
  OrbitPathSamplePoints  512
  RingSystemSections     512

  ShadowTextureSize      512
  EclipseTextureSize     256


#-----------------------------------------------------------------------


It will bring a huge improvement to the rings and orbit rendering.